Um estudo quantitativo para caracterização da qualidade interna de sistemas de software orientados a objetos Open-Source

• Ano de defesa: 2015
• Autor(a) principal: Santos, Mariana de Azevedo
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Federal de Lavras; Programa de Pós-Graduação em Ciência da Computação; UFLA; brasil; Departamento de Ciência da Computação
• Palavras-chave em Português: Qualidade de software; Software open-source; Orientação a objetos; Medidas de software; PLS-SEM; Software quality; Open-source software; Object-oriented; Software measures; Engenharia de Software
• Link de acesso: http://repositorio.ufla.br/jspui/handle/1/10561

Software development organizations are increasingly concerned with software quality assurance, given that these systems must be easily developed and maintained. However, despite necessity, activities regarding quality assurance and maintenance of software systems are considered the longest and most complex in software development lifecycle. In addition, if productivity in the development phase is low, the individuals involved in the software development process may have to invest a lot of time in post-development maintenance activities. Taking advantage of this growing trend and of the benefits obtained from open-source initiative, such as extensive knowledge exchange, automated large-scale analysis and ability to quickly drive innovations, researches on OS software quality and maintainability have gained renewed interest. The use of robust statistical techniques, such as PLS -SEM, to investigate and empirically validate software quality models has also been an efficient alternative to obtain information on OS software quality. The aim of this study was to evaluate the internal software quality by means of characterizing software domain in Java open-source systems. The study results indicate that there are: i) domains with similar traits to each other; and ii) four factors can influence the internal quality of OO software to present better maintainability, being them Reduced Complexity, Normalized Cohesion, Nonnormalized Cohesion and Increase of Modularity Degree. We propose a conceptual model to explain the relationship between the internal software quality attributes (software measures) in these factors. The results also indicate the existence of measures that can be most relevant in characterizing maintainability in OO OS software systems, such as Fan-out (FOUT), Lack of Cohesion of Methods 2 (LCOM2), Response for Class (RFC), Tight Class Cohesion (TCC), Loose Class Cohesion (LCC). Thus, this study aids software engineering professionals, allowing the development of high quality and low maintenance cost software that meet the established standards and the user needs.